The Year In AI So Far: Massive Models And How To Use Them

Ponovno objavil Platon

Spremljevalci: 0

The world of artificial intelligence and machine learning moves very fast. So fast, in fact, that it’s remarkable to think that it was only a decade ago when the AlexNet model dominated the ImageNet competition and kicked off the process that made deep learning a bona fide technology movement. Today, after years of headlines about game-playing, we see ever-increasing innovation that applies to the real world.

In the last couple of years alone, AI/ML models like GPT-3 and AlphaFold delivered capabilities that catalyzed novi izdelki in podjetja, and that stretched our understanding of what computers can do.

With that in mind, we thought we’d revisit our AI/ML coverage in Prihodnost over the first half of the year, as well as catch you up on some — but certainly not all — of the major industry developments during that time. As you’ll see, some combination of large language models, generative models, and foundation models are a major source of attention, and we’re just skimming the surface in terms of understanding what they can do and how the world outside of large research labs can utilize their power.

O Prihodnost focus: How to take advantage of AI/ML advances

How to Use Massive AI Models (Like GPT-3) in Your Startup by Elliot Turner / Hyperia

AlphaFold, GPT-3 and How to Augment Intelligence with AI by Niko Grupen / Cornell

AlphaFold, GPT-3 and How to Augment Intelligence with AI (Pt. 2) by Niko Grupen / Cornell

Data50: The World’s Top Data Startups by Jennifer Li, Sarah Wang, and Jamie Sullivan / a16z

Nastajajoče arhitekture za sodobno podatkovno infrastrukturo by Matt Bornstein, Jennifer Li, and Martin Casado / a16z

A Decade of Deep Learning: How the AI Startup Experience Has Evolved with Richard Socher (Q&A) / you.com

7 Techniques for Building Reliable AI Models by Beena Ammanath (book excerpt) /Deloitte

The Two Things We’ll Need for the Next AlphaFold with Daphne Koller (Q&A) / Insitro

Industry focus: Images, words, and more coding

Competitive Programming with AlphaCode / Deep Mind

Teaching AI to Translate 100s of Spoken and Written Languages in Real Time / Meta AI

Pathways Language Model (PaLM): Scaling to 540 Billion Parameters for Breakthrough Performance / Google Research

DALL-E2 / OpenAI

Imagen: Text-to-Image Diffusion Models / Google Research

These types of advances, and the increased understanding of how to utilize them, are why we’re dedicated to stepping up our coverage of AI/ML and, in particular, how we’ll see it applied in real-world settings over the next few years. From biotehnologije do televizija, we’re poised for a serious reimagining of what’s possible and how software can help humans deliver on their wildest ideas. If you’re working on something exciting and novel in the AI/ML space and want to share your thoughts on where we’re headed, prosim pošljite nam predlog.

Objavljeno 27. junija 2022

Tehnologija, inovacije in prihodnost, kot pravijo tisti, ki jo gradijo.

Hvala za prijavo.

Preverite svoj nabiralnik za dobrodošlico.

Stališča, izražena v »objavah« (vključno s članki, podcasti, videoposnetki in družbenimi mediji), so stališča posameznikov, ki so tam citirani, in niso nujno stališča družbe AH Capital Management, LLC (»a16z«) ali njenih povezanih družb. Nekatere informacije, vsebovane tukaj, so bile pridobljene iz virov tretjih oseb, vključno s portfeljskimi družbami skladov, ki jih upravlja a16z. Čeprav so vzeti iz virov, za katere menijo, da so zanesljivi, a16z ni neodvisno preveril takih informacij in ne daje nobenih zagotovil o trajni točnosti informacij ali njihovi ustreznosti za dano situacijo.

Ta vsebina je na voljo samo v informativne namene in se je ne smete zanašati kot pravni, poslovni, naložbeni ali davčni nasvet. Glede teh zadev se morate posvetovati s svojimi svetovalci. Sklici na katere koli vrednostne papirje ali digitalna sredstva so samo v ilustrativne namene in ne predstavljajo naložbenega priporočila ali ponudbe za zagotavljanje investicijskih svetovalnih storitev. Poleg tega ta vsebina ni namenjena nobenim vlagateljem ali bodočim vlagateljem niti ji ni namenjena in se nanjo v nobenem primeru ne smete zanašati, ko se odločate za vlaganje v kateri koli sklad, ki ga upravlja a16z. (Ponudba za vlaganje v sklad a16z bo podana le z memorandumom o zasebni plasiranju, pogodbo o vpisu in drugo ustrezno dokumentacijo katerega koli takega sklada in jo je treba prebrati v celoti.) Vse naložbe ali portfeljske družbe, omenjene, navedene ali opisane niso reprezentativne za vse naložbe v vozila, ki jih upravlja a16z, in ni nobenega zagotovila, da bodo naložbe donosne ali da bodo imele druge naložbe v prihodnosti podobne značilnosti ali rezultate. Seznam naložb skladov, ki jih upravlja Andreessen Horowitz (razen naložb, za katere izdajatelj ni dal dovoljenja a16z za javno razkritje, kot tudi nenapovedanih naložb v digitalna sredstva, s katerimi se javno trguje), je na voljo na https://a16z.com/investments/.

Grafi in grafi, ki so navedeni v informativni obliki, so zgolj informativne narave in se nanje ne smejo sklicevati pri odločitvi o naložbah. Dosedanji rezultati ne kažejo na prihodnje rezultate. Vsebina govori samo od navedenega datuma. Vse napovedi, ocene, napovedi, cilji, obeti in / ali mnenja, izražena v teh gradivih, se lahko spremenijo brez predhodnega obvestila in se lahko razlikujejo ali so v nasprotju z mnenji drugih. Prosim poglej https://a16z.com/disclosures za dodatne pomembne informacije.

Časovni žig: Junij 27, 2022Julij 16, 2022